What Do I Want from the Publisher of the Future
In this Perspective by the Editor-in-Chief of PLoS Computation Biology, the possibility of a completely integrated and public research ecosystem is explored. The author frames this in the context of a publisher, most likely due to his personal perspective as Editor-in-Chief, but a publisher-centric view is not a requirement of the analysis.
The main argument is that article publications are only one part of the production known as research and the other parts should also be published to be a part of the scholarly record. It is further argued that, with the advent of ubiquitous and cheap digital technologies, there is no excuse for not publishing those materials. The other, arguable equally or even more important, parts are grouped together under the term "workflow" and are comprised of the methods (software in some cases) and the data collected, analyzed, or used.
Another pressure to publish the workflow of research is that there is a tremendous amount of material that is easily lost due to bad information management. The author cites 4 examples from his own research (and acknowledges that he could list many more) including:
- much intellectual memory is locked away in badly managed personal email folders,
- important communication happens via presentations without a useful central location to share those materials with others,
- important pieces of software are lost due to personnel transfers, and
- data is sometimes lost for the same reason.
There are, however, issues which must be addressed. These are:
- "It is much harder to manage." Papers/Documents/PDFs are a known and mostly solved problem within the area of scholarly communication. A dynamic workflow, is not.
- "The scientific endeavor as a simple linear workflow is also clearly an oversimplification." Managing distinct steps along a known path is easy. But managing the various and multiple paths taken during a process is much harder, especially when many steps are undertaken multiple times with differing parameters.
- Confidentiality: Some research will need to be private during the process due to either personal opinions on when it is acceptable to share or due to research participant privacy issues. Philip speculates that when given the choice, more scientific information will be shared earlier within this system.
- Peer Review: It would be important for expert peers to also review the workflow of a researcher, not just their submitted paper, and in some progressive cases, their supporting datasets.
- New Infrastructure: There are already systems in place for managing the peer-review of articles but there needs to be a similar solution for the review of workflows.
- Data Repositories vs Publishers: Publishers do not inherently need to host all materials associated with a publication and can instead provide strong interoperability options with data repositories.
- Community Acceptance: With a seemingly radical jump in the services provided by a publisher this may cause some members of the scholarly community to be hesitant to accept the new paradigm. However, some publishers already provide some services to accept and publish additional parts of the workflow than just the final article now, and extending those offerings seems logical. Additionally, an acceptance of this new material within the reward system of academia is a crucial step.
- Journals and the Reward System: Scientific success has largely been associated with (seemingly) simple metrics associated with only article publications. Adding other materials from the research workflow is just another facet to be considered and its value measured.
If all of these issues are addressed and the new system is accepted then, "the complete result will be a digital workflow that begins with a documented idea and ends in a set of conclusions from a scientific experiment, all of which will be published by the publisher of the future and accepted as the norm in scholarly communication."